Exploring the Merits of Commercial Vehicle Forward Collision Avoidance
and Mitigation Systems (F-CAM) Presented to:
Task Force on Vehicle Weights and Dimensions Policy
John Woodrooffe
November 20, 2013
Montreal
Private Public Partnership Sponsor
NHTSA and Meritor WABCO NHTSA – Robert Kreeb and James Funke Meritor WABCO – Alan Korn (Jon Morrison)
2
Terminology System is comprised of Forward Collision
Warning + Autonomous Breaking Forward collision contributes to crash
avoidance and autonmated braking contributes to collision mitigation Commercial Vehicle Forward Collision
Avoidance and Mitigation Systems (F-CAM)
3
Project Goals
Characterize the performance of a current F-CAM system via test track experiments and simulation.
Identify and profile the target crash population for F-CAM systems (i.e. truck-involved rear-end crashes).
Estimate, via modeling and simulation, the effectiveness of F-CAM technologies in avoiding and mitigating rear-end crashes
Obtain “case and control” data from fleets for statistical analysis of F-CAM safety performance in real-world application.
Apply cost factors to crash reduction/mitigation estimates to determine total economic benefits
4
Project Elements
Crash data analysis Fleet data (2 national fleets analyzed) Test program Modeling Benefit analysis
5
F-CAM Intervention Sequence
t0
Object tracked
Collision warning:
Visual and Audible
Collision warning:
Haptical (short brake pulse)
Automatic braking for
collision prevention or
mitigation
Avoidance maneuver not
possible
time t2 t3 t4
Engine Torque Limitation Brake Activation
Potential rear end collision
detected
Hard braking required to
prevent collision
t1
Warning Tone and Lamp System Reactions
Crash prevented
or mitigated
6
7
Crash types selected as relevant to the technology
Rear-end, striking Current generation: Lead vehicle stopped at impact, but seen moving Lead vehicle slower, steady speed Lead vehicle decelerating Lead vehicle cut-in
Target Crash Types
8
Estimated Annual Rear-end Striking Crashes TIFA 2003-2008, GES 2003-2008
“Fixed” means LV was stationary (fixed) before coming in radar range of the subject vehicle, i.e., never seen moving.
“Stopped” means LV seen moving by the subject vehicle’s radar prior to coming to a stop.
Tractor Semitrailer
9
Crash type
Fatal Injury PDO Total N N N N
LV fixed 62 882 2,119 3,078 LV stopped 13 1,244 2,987 4,263
LV slower 90 1,199 1,794 3,082 LV decel. 18 1,502 3,152 4,750
LV cut-in 9 156 649 814
Total 192 4,983 10,701 15,987* “PDO” specifies property damage only crashes. * Total includes 111 crashes of unknown injury severity.
Single Unit Truck
Crash type Fatal Injury PDO Total
N N N N
LV fixed 20 1,215 2,202 3,438
LV stopped 8 2,228 4,037 6,270 LV slower 26 318 902 1,246 LV decel. 8 1,222 3,815 5,096 LV cut-in 1 134 187 322 Total 63 5,117 11,143 16,374*
Fatalities and Injuries in Rear-end Striking Crashes TIFA 2003-2008, GES 2003-2008
“Fixed” means LV was stationary (fixed) before coming in radar range of the subject vehicle, i.e., never seen moving.
“Stopped” means LV seen moving by the subject vehicle’s radar prior to coming to a stop.
Tractor Semitrailer
10
Single Unit Truck
Crash type
Injury severity Total injuries Fatal A-
injury B-injury C-injury
LV fixed 78 139 335 861 1,413 LV stopped 16 158 431 1,179 1,782
LV slower 107 601 865 727 2,300
LV decelerating 22 303 605 1,251 2,180
LV cut-in 9 87 48 115 259
Total 231 1,287 2,284 4,132 7,934
Crash type
Injury severity Total injuries Fatal A-
injury B-
injury C-
injury LV fixed 22 156 278 1,272 1,728 LV stopped 9 277 493 2,306 3,085 LV slower 30 116 154 241 542 LV decelerating 10 189 334 1,426 1,959 LV cut-in 1 2 38 141 182 Total 72 740 1,298 5,386 7,496
Subject Vehicle Highlights
SV Brush Guard UMTRI DAS
Forward Radar Multiple DVI DAS Interface
11
Towable Target Evolution and Highlights
“Seed” Initial UMTRI Radar only Target
Initial Vision Compatible Target
Vision Compatible Target
Final Vision Compatible Target
12
Establishing the Simulated “Reference” (or baseline) Crash Database
13
IVBSS Heavy Truck Feld Operational Test
1,000,000 km by 18 drivers over 10 months Mix of P&D and Line-haul
Initial Conditions (Speeds, Distance, PovAx) Lead Vehicle Braking, N = 8210 events Lead Vehicle Slower, N = 1471 events Cut-in, N = 382 events Fixed, N = 470 events
Driver Braking Profile Driver Brake Reaction Time
14
Effect of Delay Time on Severity 60
50
40
30
20
10
0
Ran
ge, m
-14 -12 -10 -8 -6 -4 -2 0Range-rate, m/s
Increasing Delay Time
Driver Delay Time to Braking, 0 s
60
50
40
30
20
10
0
Ran
ge, m
-14 -12 -10 -8 -6 -4 -2 0Range-rate, m/s
Increasing Delay Time
Driver Delay Time to Braking, 0 s Driver Delay Time to Braking, 1 s
60
50
40
30
20
10
0
Ran
ge, m
-14 -12 -10 -8 -6 -4 -2 0Range-rate, m/s
Increasing Delay Time
Driver Delay Time to Braking, 0 s Driver Delay Time to Braking, 1 s Driver Delay Time to Braking, 2 s Driver Delay Time to Braking, 3 s
60
50
40
30
20
10
0
Ran
ge, m
-14 -12 -10 -8 -6 -4 -2 0Range-rate, m/s
Increasing Delay Time
Driver Delay Time to Braking, 0 s Driver Delay Time to Braking, 1 s Driver Delay Time to Braking, 2 s Driver Delay Time to Braking, 3 s Driver Delay Time to Braking, 4 s
60
50
40
30
20
10
0
Ran
ge, m
-14 -12 -10 -8 -6 -4 -2 0Range-rate, m/s
Increasing Delay Time
Driver Delay Time to Braking, 0 s Driver Delay Time to Braking, 1 s Driver Delay Time to Braking, 2 s Driver Delay Time to Braking, 3 s Driver Delay Time to Braking, 4 s Driver Delay Time to Braking, 4.2 s
60
50
40
30
20
10
0
Ran
ge, m
-14 -12 -10 -8 -6 -4 -2 0Range-rate, m/s
Increasing Delay Time
Driver Delay Time to Braking, 3 s Driver Delay Time to Braking, 4 s Driver Delay Time to Braking, 4.2 s Driver Delay Time to Braking, 4.3 s
Increasing Crash Severity
60
50
40
30
20
10
0
Ran
ge, m
-14 -12 -10 -8 -6 -4 -2 0Range-rate, m/s
Increasing Delay Time
Driver Delay Time to Braking, 3 s Driver Delay Time to Braking, 4 s Driver Delay Time to Braking, 4.2 s Driver Delay Time to Braking, 4.3 s Driver Delay Time to Braking, 6.8 s
Crash
Initial Conditions: Sv Speed = 40 mph Pov Speed = 38 mph Range = 55 m Pov Decel Profile
-4
-3
-2
-1
0
Acc
eler
atio
n, m
/s2
43210Time, s
Ax Profile for Driver Braking
Acceleration Profile used for Driver Braking
Sv Decel Profile
15
Estimating delta-V distribution for historical rear-end crash population
We get baseline delta-V distribution by finding the distribution of delta-V that reproduces the injury patterns for truck-into-car rear ends in GES
X
Delta-V (Exposure) X Risk (given dV)
=
= Injury
Unique delta-V distributions are developed for each crash type (LV slower, decelerating, stopped, cut-in).
16
Evaluate system performance compared to baseline
System evaluation: 1) FCW—accounts for a distribution of driver
brake reaction times from 0.5-2.7 sec, based on literature and braking in IVBSS
2) CMB—three systems; driver not in loop. 3) Combination—best performance of either
FCW or CMB for each case
17
Technology Simulation Methodology
Based on the rules for a FCW, calculate the simulation time when an FCW would have been given to the driver
Map Driver Brake Reaction Time Distribution on to Baseline Simulations
For each Baseline simulation that resulted in a crash—rerun with the three CMB algorithms and save the results
18
Characteristics of Future Systems
System can reliably detect moving and fixed vehicles
CMB automated braking deceleration levels nominal 0.35 g for the second generation system nominal 0.60 g for the third generation system
19
Reduction in Injury Severity
20
Tractor Semitrailer Single Unit Trucks Device Fatal Injury No
injury
Subsystem Contribution
FCW only 31% 27% 11%
CMB only 2nd gen. 26% 32% 10%
CMB only 3rd gen. 44% 42% 19%
Complete System Contribution Second
Generation 44% 47% 20%
Third Generation 57% 54% 29%
Current Generation 24% 25% 9%
Device Fatal Injury No injury
Subsystem Contribution
FCW only 28% 25% 11%
CMB only 2ND Gen. 27% 33% 13%
CMB only 3rd Gen. 42% 46% 23%
Complete System Contribution Second
Generation 43% 46% 24%
Third Generation 55% 57% 34%
Current Generation 22% 21% 10%
Total Annual Economic Benefit (2013 Dollars)
21
Tractor Semitrailer Single Unit Trucks Device Fatal Injury No
injury Total
Subsystem Contribution
FCW only $528.9 $544.8 $34.4 $1,108.1
CMB only 2nd gen. $446.2 $633.6 $31.9 $1,111.7
CMB only 3rd gen. $741.2 $792.8 $60.6 $1.594.6
Complete System Contribution Second
Generation $745.0 $919.5 $65.8 $1,730.3
Third Generation $972.7 $1046.1 $93.1 $2,112.0
Current Generation $412.4 $513.0 $29.5 $954.9
Device Fatal Injury No injury Total
Subsystem Contribution
FCW only $142.3 $395.3 $30.5 $568.1
CMB only 2nd Gen. $134.6 $500.8 $35.4 $670.8
CMB only 3rd Gen. $211.7 $690.2 $62.4 $964.3
Complete System Contribution Second
Generation $214.7 $703.8 $63.9 $982.4
Third Generation $275.6 $853.9 $89.7 $1,219.2
Current Generation $112.9 $342.8 $25.8 $481.5
Conclusions Tractor semitrailers The annual reduction in fatalities and injuries relative to the base
population for current generation systems is: Current technology 24% and 25% respectively ($0.9 billion/yr) Second generation 44% and 47% respectively ($1.7 billion/yr) Third generation 57% and 54% respectively ($2.1 billion/yr) Single Unit Trucks The annual reduction in fatalities and injuries relative to the base
population for current generation systems is: Current technology 22% and 21% respectively ($0.5 billion/yr) Second generation 43% and 46% respectively ($1.0 billion/yr) Third generation 55% and 57% respectively ($1.2 billion/yr)
22
Conclusions
Current generation F-CAM systems provide significant reduction in the frequency and severity of truck rear-end striking crashes
The research indicates that future systems will provide additional benefit:
Second generation – factor 1.9 Third generation – factor of 2.3
(relative to current generation systems)
23
Thank You! [email protected]
24